Development and Validation of Inverse Model to Detect Fire Source and Intensity.

A model and procedure to detect fire location and inverse fire intensity is developed. Markov Chain Monte Carlo sampling based on the Bayesian inference is used to invert the parameters such as source location and its strength. Two test cases are used to evaluate the model. First, the model is validated using experimental data from the “NBS Multi-room Test Series”. Second, a two-story office building fire with 35 compartments is used to investigate the sensitivity and reliability of the model. It is shown that predicted fire source and intensity agree well with the actual value. Then the effects of the sensors’ time sampling interval and intersensor spacing on the sensitivity and reliability of the method are studied respectively. The results indicate that small time sampling interval generally result in high estimation performance, but the decreasing of the intersensor space is not significantly helpful to improve the accuracy of the inverse intensity if the time sampling interval is small enough. In addition, it is discovered that the accuracy of the predicted fire location is not affected by the accuracy of the forward fire model, while the accuracy of predicted fire intensity is sensitive to the systematic errors or the accuracy of the forward model. [ABSTRACT FROM AUTHOR]